There is substantial evidence that genetic events are associated with the etiology of brain tumors. The purpose of this study is to characterize the cancer risk in a large number of kindreds of glioma patients using combined epidemiologic and statistical techniques and incorporating molecular genetic markers. We will continue to accrue families to obtain a sufficiently large sample size to have precise estimates of the heritable component of gliomas. Several types of analyses will be performed on the data set. First, a descriptive analysis to detect overall cancer excess or excess of specific sites by computing Standardized Incidence Ratios (SIRs). The SIRs will determine whether excess cancer exists in first-degree relatives of glioma patients, by specific site, or by specific proband or relative characteristics. The next type of analysis is hypothesis testing using segregation analysis (with the computer program, Statistical Analysis for Genetic Epidemiology - SAGE) to determine the genetic model and parameter estimates that characterize the mode of inheritance of gliomas and identify specific kindreds showing evidence for a major gene to include in linkage analysis. Through linkage candidate gene analysis we will determine if mutations at distinct cancer predisposing loci are etiologically relevant and inherited from an antecedent relative. The candidate gene analysis will be conducted by Dr. Peter O'Connell at The University of Texas Health Science Center at San Antonio. With a larger sample we will be able to confirm our finding that patients with multifocal glioma, gliomas as a second malignancy, and a family history constitute a genetically predisposed subgroup. This study will be the first comprehensive evaluation of familial aggregation in families of unselected glioma patients. In addition, we are collecting blood from each new patient and our specimen repository will be available to study future candidate genes to integrate with the family data.